New Ryzen 2 (Pinnacle Ridge) gets only 200 MHz boost according to a leak

B
mkrohn said:
Ryzen is great for mining but not much else. just like the vega 64 i have paired with my Ryzen 1700. I'm about to buy an 8700K rig with a 1080ti in it since thats the only reasonable route to buy a 1080ti right now. pre builts are about the only way to buy a 1080ti right now. I've been holding out solid news on the newer chips first. It looks like a safe time to jump back to intel. Thanks for the price drops AMD but you're not gonna be in my desktop.
Click to expand...

Because there is now only 2 uses for desktops:
Ultra high fps 1080p gaming and mining, right.
 
Last edited:
juanrga said:
Virtually everyone in the tech news is referring to Pinnacle Ridge as Ryzen 2

https://www.google.com/search?q="Ryzen+2"

Also "Ryzen 2" is routinely used here in the forums

https://hardforum.com/search/10909932/?q="Ryzen+2"&o=relevance

So people is using

Ryzen 1 = Summit Ridge, Zen @14nm

Ryzen 2 = Pinnacle Ridge, Zen @14nm+ (renamed to Zen+ and 12nm)

If AMD finally comes with a different commercial name than Ryzen 2 I don't know.
Click to expand...

Skylake is 6th generation Intel Core processor

Zen+ is 2nd generation AMD Ryzen processor

If somebody spelled your name wrong, you name didn't suddenly changed.
 
If Zen+ performs 5% better at same clock and max achievable clock is now 10% higher (4,4 ~ 4,5GHz), it is already a huge win. It is worth considering that AMD seems ambitious in selling all of their current stock. They're probably doing this because if Zen+ is a significant improvement, they don't want all the Ryzen first generation CPUs collecting dust on the shelves for eternity like Bulldozer did.
 
Briguyatx said:
I am still a bit salty about my "upgrade" from a rock solid 2700k at 4.8 to a 4770k that gets pissy if you try to run 4.3 . I must admit, the 4770K works just fine with a 1080ti 4 years later, and I doubt the 2700k would still be this healthy.

Upgrade itch, must scratch, It's been entirely too long, and 5+ghz on an 8700k sounds nice. I wanted a 1700x but the single thread perf isn't much of an improvement and the game I play the most is basically constrained to no more than 2 cores, and still really only stresses one (World of Warcraft).

I'm hoping the zen+ 1700x equivalent can get up to 4.5, if so I will be sold on it.
Click to expand...


I am in the same boat, Hoping for the day WoW makes use of more than a single core or two. But, I went the route of AMD 1700, and just OC to 3600. I have not had any issue in WoW, and while I was running EZ CD last night and converting 3900 songs from my library and playing with no slowdowns in WoW, sure made me smile :)
 
According to a AMD guy on reddit, the Ryzen Mobile are technically 14nm+, which are basically offering savings in power, enough to create Precision Boost 2. He notes hes seen a Ryzen 5 2400G running at 4.2GHz with the Wraith Max (a Hyper 212 EVO equivalent), and at CES, they demoed it running 3.9Ghz with all cores with DDR4-3600, and the GPU clocked to 1675Mhz vs base 1250Mhz. Now hes very careful to note that it was a production CPU, he has no data or knowledge if this is nothing more than a cherry-picked golden sample.
 
SighTurtle said:
According to a AMD guy on reddit, the Ryzen Mobile are technically 14nm+, which are basically offering savings in power, enough to create Precision Boost 2. He notes hes seen a Ryzen 5 2400G running at 4.2GHz with the Wraith Max (a Hyper 212 EVO equivalent), and at CES, they demoed it running 3.9Ghz with all cores with DDR4-3600, and the GPU clocked to 1675Mhz vs base 1250Mhz. Now hes very careful to note that it was a production CPU, he has no data or knowledge if this is nothing more than a cherry-picked golden sample.
Click to expand...

It makes sense. In fact innovations in Zen+ such as "Precision Boost 2" aren't new but already present in Zen mobile and desktop products and advertised in AMD site since past year.

According to official data the Ryzen 5 2400G has a max boost of 3.9GHz but XFR is unknown. Maybe 4.2GHz is XFR. In this case we can expect about 300MHz extra for XFR in Pinnacle Ridge.

EDIT:

He also explains the need for the 400-series chipset: "The 400 Series motherboards are designed to improve memory stability and overclocking."
 
Last edited:
Not sure that I am following. I though all the new APUS were still not Pinnacle based. The may have added precision boost 2 but I thought they are still 14 nm.

Sorry if I misunderstood
 
juanrga said:
It makes sense. In fact innovations in Zen+ such as "Precision Boost 2" aren't new but already present in Zen mobile and desktop products and advertised in AMD site since past year.
Click to expand...

Precision Boost 2 is new.

It was launched with Raven Ridge.

Summit Ridge doesn't have it.
 
_mockingbird said:
Precision Boost 2 is new.

It was launched with Raven Ridge.

Summit Ridge doesn't have it.
Click to expand...

I didn't mention Summit Ridge. When I mentioned "Zen mobile and desktop products" I did mean APUs. So, as stated in my former post, precision boost 2 is not a new technology introduced only on Zen+ cores. Precision boost 2 was launched with Raven Ridge, which uses Zen cores.

In fact all the known innovations associated to Zen+ products (precision boost 2, 14nm+, faster IMC) are already present in Zen products (Raven Ridge).
 
juanrga said:
I didn't mention Summit Ridge. When I mentioned "Zen mobile and desktop products" I did mean APUs. So, as stated in my former post, precision boost 2 is not a new technology introduced only on Zen+ cores. Precision boost 2 was launched with Raven Ridge, which uses Zen cores.

In fact all the known innovations associated to Zen+ products (precision boost 2, 14nm+, faster IMC) are already present in Zen products (Raven Ridge).
Click to expand...

Hence the 2xxx naming scheme for the apu's, they're basically Zen+ cores.
 
juanrga said:
It makes sense. In fact innovations in Zen+ such as "Precision Boost 2" aren't new but already present in Zen mobile and desktop products and advertised in AMD site since past year.

According to official data the Ryzen 5 2400G has a max boost of 3.9GHz but XFR is unknown. Maybe 4.2GHz is XFR. In this case we can expect about 300MHz extra for XFR in Pinnacle Ridge.

EDIT:

He also explains the need for the 400-series chipset: "The 400 Series motherboards are designed to improve memory stability and overclocking."
Click to expand...

That seems to me incorrect. The memory compatibility should have next to nothing to do with the chipset. It is the IMC in the processor that determines memory compatibility and the bios enables that feature. The x470 chipsets merely convert previous pciE 2.0 lanes to pciE 3.0 lanes. Theoretically an x370 chipset with an updated Pinnacle Ridge bios should allow improved meory compatibility if you install a Pinnacle Ridge cpu. My thought is that the bios may not allow that so that motherboard makers can sell more motherboards with no overriding technical issue to really justify that action, only greed.
 
os2wiz said:
That seems to me incorrect. The memory compatibility should have next to nothing to do with the chipset. It is the IMC in the processor that determines memory compatibility and the bios enables that feature. The x470 chipsets merely convert previous pciE 2.0 lanes to pciE 3.0 lanes. Theoretically an x370 chipset with an updated Pinnacle Ridge bios should allow improved meory compatibility if you install a Pinnacle Ridge cpu. My thought is that the bios may not allow that so that motherboard makers can sell more motherboards with no overriding technical issue to really justify that action, only greed.
Click to expand...

The DIMM modules don't connect directly to the CPU. The signals from the modules to the CPU and viceverse have to travel through the circuitry and wires in the actual motherboard. If the motherboard cannot produce fast enough signals or if those signals aren't stable, then it matter little how many improvements you make to the CPU.

The AMD guy is clever: "400 Series motherboards are designed to improve memory stability and overclocking". So, by design, you don't get those improvements in older 300-series mobos.
 
bobzdar said:
Hence the 2xxx naming scheme for the apu's, they're basically Zen+ cores.
Click to expand...

You really mean Zen cores ;)

csm_ryzen_apu_pres_10_e6138a59bb.jpg


RyzenMobile-Data1-Large.jpg


Those 2000 series APUs have Zen cores inside.
 
Correct, Juangra.

Shame on AMD for using the 2000 series name for a cpu with zen cores and not zen+. Hardware Canucks mentioned this as well.
 
juanrga said:
You really mean Zen cores ;)

csm_ryzen_apu_pres_10_e6138a59bb.jpg


RyzenMobile-Data1-Large.jpg


Those 2000 series APUs have Zen cores inside.
Click to expand...


juanrga said:
I didn't mention Summit Ridge. When I mentioned "Zen mobile and desktop products" I did mean APUs. So, as stated in my former post, precision boost 2 is not a new technology introduced only on Zen+ cores. Precision boost 2 was launched with Raven Ridge, which uses Zen cores.

In fact all the known innovations associated to Zen+ products (precision boost 2, 14nm+, faster IMC) are already present in Zen products (Raven Ridge).
Click to expand...

So what's the difference between a zen+ core and a zen core with all of the zen+ improvements, exactly?
 
Nightfire said:
Correct, Juangra.

Shame on AMD for using the 2000 series name for a cpu with zen cores and not zen+. Hardware Canucks mentioned this as well.
Click to expand...

The shame was renaming Zen to Zen+.

bobzdar said:
So what's the difference between a zen+ core and a zen core with all of the zen+ improvements, exactly?
Click to expand...

The difference is that one of them was re-branded for marketing purposes.
 
Last edited:
Zen+ hasn't even been tested yet. If i hits at least 4.4 ghz all core when o/c, I will be happy with them calling it whatever they want.

Anything below that will be a fail since a cream of the crop 1900x can do 4.3 ghz all core.
 
How long until Ryzen 2? I'll probably end up skipping this because I don't see it being worth the time upgrading from my Ryzen 1600, but I'm eagerly awaiting Ryzen 2.
 
Rakanoth said:
https://hothardware.com/news/amd-ryzen-5-2600-12nm-zen-cpu-asus-crosshair-vii-hero-x470-motherboard

If this leak is true, I will stay away from AMD forever and ever. I will sell my Ryzen 1700 right away. Sorry AMD.
Click to expand...


Fake news by a headline seeker. You damn well know this is an engineering sample and silicon has not been finalized. It will be at least 400MHZ improvement for the 2800X which comes to a 10% ramp up in clock speed. The cache latency improvement will up the IPC by about 3-4%. That means over a 10% performance improvement with no increase in thermals. I consider that a major success for a minor process improvement. That makes the product the equal or better than Kaby Lake and CoffeeLake at a better price point.
 
juanrga said:
The AMD guy is clever: "400 Series motherboards are designed to improve memory stability and overclocking". So, by design, you don't get those improvements in older 300-series mobos.
Click to expand...

The Intel guy is clever: "300 Series motherboards are designed to improve memory stability and overclocking". So, by design, you don't get those improvements in older 200-series mobos.

Funny how that works. At least with AMD you get socket compatibility between generations, and they aren't spreading the gospel of "more pins for better power delivery" to the masses.
 
Last edited:
juanrga said:
The DIMM modules don't connect directly to the CPU. The signals from the modules to the CPU and viceverse have to travel through the circuitry and wires in the actual motherboard. If the motherboard cannot produce fast enough signals or if those signals aren't stable, then it matter little how many improvements you make to the CPU.

The AMD guy is clever: "400 Series motherboards are designed to improve memory stability and overclocking". So, by design, you don't get those improvements in older 300-series mobos.
Click to expand...

Talking about those new low latency traces on the motherboard right? Super voltage regulators? Or just updated bios?
 
juanrga said:
The shame was renaming Zen to Zen+.



The difference is that one of them was re-branded for marketing purposes.
Click to expand...


So it's zen+ then. FFS you have to obfuscate everything, don't you? They changed the imc, improved l3 cache and did a node shrink. That's more than intel usually does when going up a generation.
 
os2wiz said:
Fake news by a headline seeker. You damn well know this is an engineering sample and silicon has not been finalized.
Click to expand...

The chip leaked is not an engineering sample. It is a qualification sample.

bobzdar said:
So it's zen+ then. FFS you have to obfuscate everything, don't you? They changed the imc, improved l3 cache and did a node shrink. That's more than intel usually does when going up a generation.
Click to expand...

You insist on that the 2000 series APUs use zen+ cores, but I gave you slides from AMD stating clearly those APUs use "Zen cores".

* The IMC is not part of the core. And it simply operating at higher frequency now. 2666MHz ---> 2993MHz

* The L3 cache is not part of the core. And again the L3 cache is operating at higher frequencies now.

* A node shrink doesn't change the core microarchitecture. Moreover 12LP is not a node shrink. It is a relabel of 14nm+.

So the cores on the 2000-series APUs and in Pinnacle Ridge are the same: Zen cores. The only difference is AMD is using a more mature process node (14nm+) to get higher clocks than Summit Ridge (14nm).
 
Last edited:
juanrga said:
The chip leaked is not an engineering sample. It is a qualification sample.



You insist on that the 2000 series APUs use zen+ cores, but I gave you slides from AMD stating clearly those APUs use Zen cores.

* The IMC is not part of the core. And I don't think they changed the IMC, simply it is operating at higher frequency now. 2666MHz ---> 2933MHz

* The L3 cache is not part of the core. And the improvement on the L3 cache comes from higher frequencies by virtue of IF.

* A node shrink doesn't change the core microarchitecture. Moreover 12LP is not a node shrink. It is a relabel of 14nm+.

So the cores are the same: Zen cores. The only difference is using a more mature process node to get higher clocks.

The same happened with Trinity and Richland or with Kaveri and Godavari. Same cores (Piledriver in one case, Steamroller in the other) and the difference between generations was higher clocks thanks to a more mature process (32nm in one case, 28nm in the other).
Click to expand...

The imc and cache are part of the zen architecture. They can't be decoupled.

What is the difference between the apu architecture and zen+? You insist this is not the same as zen+ but cannot explain the difference between the apu architecture with changed imc, reduced latency L3, node shrink and 2xxx naming scheme and the zen+ that is exactly the same thing (except with two 4 core complexes).
 
bobzdar said:
The imc and cache are part of the zen architecture. They can't be decoupled.

What is the difference between the apu architecture and zen+? You insist this is not the same as zen+ but cannot explain the difference between the apu architecture with changed imc, reduced latency L3, node shrink and 2xxx naming scheme and the zen+ that is exactly the same thing (except with two 4 core complexes).
Click to expand...

Explained above that IMC and L3 cache aren't part of the Zen cores. Also explained IMC and the L3 cache are simply working at higher frequencies now.

Older roadmaps showed that Pinnacle Ridge uses Zen cores on 14LPP+ node. Newest roadmaps rebranded Zen to Zen+ and rebranded 14LPP+ to 12LP only for marketing purposes.

There is no node shrink. Even Globalfoundries Chief admits 12nm is only an extension of 14nm.
 
Last edited:
juanrga said:
I will explain this one time more.

The IMC and the L3 cache are working at ~10% higher clocks on Raven Ridge and Pinnacle Ridge. Also IMC and L3 aren't part of the core (even one of the slides I gave you show the L3 outside the "Zen cores"). So even if engineers were to change the architecture of the L3 and the IMC, the core would be the same. But as stated enginers didn't change anything, both IMC and L3 are simply running at higher clocks.

The 2000 series APUs use Zen cores (again this is in the slides that I gave you and repeated by AMD's James). Pinnacle Ridge also uses Zen cores, there is no fundamental difference between the cores used in those 2000 series APUs and the cores used in Pinnacle Ridge.

Older roadmaps showed that Pinnacle Ridge uses Zen cores on 14LPP+ node. Newest roadmaps rebranded Zen to Zen+ and rebranded 14LPP+ to 12LP only for marketing purposes.

There is no node shrink. Even Globalfoundries Chief admits 12nm is only an extension of 14nm (aka what some of us denote as 14nm+). A true shrink to 12nm would bring 36% higher density, which of course is not the case with 12LP which is only a marketing rebrand.
Click to expand...

So, I'll ask for the 3rd time, what's the difference between what is in the 2xxx APU's and what is coming with the zen+ ryzen 2xxx processors?
 
bobzdar said:
So, I'll ask for the 3rd time, what's the difference between what is in the 2xxx APU's and what is coming with the zen+ ryzen 2xxx processors?
Click to expand...

How many times I have to say the same? There is no known differences except clocks and that RR uses CCX with half-cache mode.
 
You must log in or register to reply here.
Back
Top